Action-relation hallucination evaluation on MMRel DALL-E
75.23AccuracyRVE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| RVELVLM Model=LLaVA-1.5-13B2026.05 | 75.23 | 78.31 | |
| RVELVLM Model=InstructBLIP-7B2026.05 | 72.78 | 75.64 | |
| RVELVLM Model=ShareGPT4V-7B2026.05 | 72.53 | 75.62 | |
| VanillaLVLM Model=LLaVA-1.5-13B2026.05 | 72.44 | 77.55 | |
| ICDLVLM Model=LLaVA-1.5-13B2026.05 | 71.77 | 77.19 | |
| RVELVLM Model=LLaVA-1.5-7B2026.05 | 71.6 | 75.07 | |
| VCDLVLM Model=InstructBLIP-7B2026.05 | 71.09 | 75.43 | |
| VAFLVLM Model=LLaVA-1.5-13B2026.05 | 70.84 | 77.02 | |
| RVELVLM Model=LLaVA-NeXT-7B2026.05 | 70.58 | 75 | |
| VCDLVLM Model=ShareGPT4V-7B2026.05 | 70.16 | 75.05 | |
| VanillaLVLM Model=InstructBLIP-7B2026.05 | 69.65 | 75.19 | |
| ICDLVLM Model=LLaVA-1.5-7B2026.05 | 69.57 | 72.35 | |
| VanillaLVLM Model=LLaVA-1.5-7B2026.05 | 69.06 | 73.28 | |
| VanillaLVLM Model=ShareGPT4V-7B2026.05 | 69.06 | 74.86 | |
| ICDLVLM Model=ShareGPT4V-7B2026.05 | 68.81 | 74.53 | |
| VAFLVLM Model=InstructBLIP-7B2026.05 | 68.47 | 75.31 | |
| VAFLVLM Model=ShareGPT4V-7B2026.05 | 67.79 | 74.31 | |
| ICDLVLM Model=LLaVA-NeXT-7B2026.05 | 67.62 | 74.79 | |
| VAFLVLM Model=LLaVA-1.5-7B2026.05 | 67.54 | 73.84 | |
| VanillaLVLM Model=LLaVA-NeXT-7B2026.05 | 67.54 | 74.67 | |
| VCDLVLM Model=LLaVA-1.5-13B2026.05 | 67.54 | 75.26 | |
| ICDLVLM Model=InstructBLIP-7B2026.05 | 67.29 | 73.44 | |
| VAFLVLM Model=LLaVA-NeXT-7B2026.05 | 66.78 | 74.36 | |
| VCDLVLM Model=LLaVA-1.5-7B2026.05 | 62.97 | 72.45 | |
| VCDLVLM Model=LLaVA-NeXT-7B2026.05 | 62.46 | 72.86 |